中文摘要：

支持向量机（SVM）是在结构风险最小化的一种新的机器学习技术，在解决小样本、非线性及高维空间问题中具有独特的优势，适用于政府采购中对供应商进行信用分析。但供应商信用属性数据构成了高维空间的稀疏分布，不利于SVM的准确建模。由于主成分分析技术具有良好的去噪音特性，能够对信用属性数据进行有效地挖掘。因此，若将两者进行有机地结合，就能有效改善SVM输入样本的特性，从而提高SVM分类的准确率。

英文摘要：

Support Vector Machines （SVM） based on structural risk minimization （SRM） principle is a new machine learning technique and has many advantages in solving small sample size, nonlinear and high dimensional pattern recognition. In this paper, it is applied to the credit scoring prediction of suppliers in the government procurement activities. To get better classification accuracy, PCA （ Principal Component Analysis ） is combined to SVM to mine the independent attributes of supplier credit. And then, SVM is trained by these independent attributes obtained. By this way, the model of PCA - SVM for credit ananlysis of suppliers in the government procurement activities is built to evaluate the prediction accuracy of PCA - SVM, while comparing its performance with those of neural networks （ NN ） and traditional SVM.